Computational algorithms for extraction and analysis of two-dimensional transportation networks

Werner Baumgarten

2012 · 10 citations · 16 references

Concepts

Abstract

Transportation networks are of paramount importance in natural and man-made systems. We present algorithms to extract twodimensional real world networks from images, thus making the networks available for in silico analysis. The algorithms deal with image processing, network extraction, and correction of reconstruction errors. Next, methods are presented to obtain the geometrical and topological structure of the network. This yields the graph underlying the network, which can be subjected to graph theoretical analysis. The statistical distributions of the lengths, widths, and areas of the vein segments, as well as the angles between veins departing from a branching point are investigated. In a case study, the algorithms are applied to the tubular vein network of the myxomycete Physarum polycephalum . Despite of the large variability between different network realizations, statistics obtained from cummulated experiments correspond to those extracted from individual experiments.

References

16